
# Here, we will run the main model 10 times, adding in each indicator cumulatively.
# We will loop through K [1,10] and store the 2005/2009 and 2017 results for each K.
# This takes a long time...

rm(list=ls())
## LOAD PACKAGES
require(sandwich)
require(lmtest)
require(lfe)
require(Hmisc)
require(data.table)
require(gtools)
require(tidyverse)
require(stargazer)
require(reshape)


# LOOP THROUGH 2005/2009 sample, and 2017 sample

## LOAD DATA
list_of_plot_data = list()
for(k in 1:10){
###################################
#### 2005/2009 SAMPLE ANALYSIS ####
###################################
load("02-data/befm-main-analysis.Rdata")
  df = as.data.table(df)
  df = df[(df$state.vfile == 'CA05' & age <=35) |(df$state.vfile == 'NC09' & age <=31)]
  
##############################
####### MAKE VARIABLES #######
####### FOR ANALYSIS   #######
##############################


df$white = ifelse(df$race == 100, 1, 0) # WHITE INDICATOR

#### FIXED EFFECT VARIABLES

# county
df$county_state = paste0(df$statefip, "_", df$county)  


# enumdist
df$cnty_enumdist = paste0(df$county_state, "_", df$supdist,"_",df$enumdist)
df$cnty_enumdist = ifelse(grepl('NA', df$cnty_enumdist), NA, df$cnty_enumdist)
# 10 page
df$page10 = round(df$pageno, digits = -1)
df$reel_page10 = paste0(df$reel, "_", df$page10)

# 5 page
mround <- function(x,base){ 
  base*round(x/base) }
df$page5 = mround(df$pageno, 5)
df$reel_page5 = paste0(df$reel, "_", df$page5)


df$hs_dist1 = ifelse((df$hs_opposite1_dist_1 == 1 & df$hs == 0) | (df$hs_opposite1_dist_1 == 0 & df$hs == 1),1,0)
df$hs_dist2 = ifelse((df$hs_opposite1_dist_2 == 1 & df$hs == 0) | (df$hs_opposite1_dist_2 == 0 & df$hs == 1),1,0)
df$hs_dist3 = ifelse((df$hs_opposite1_dist_3 == 1 & df$hs == 0) | (df$hs_opposite1_dist_3 == 0 & df$hs == 1),1,0)
df$hs_dist4 = ifelse((df$hs_opposite1_dist_4 == 1 & df$hs == 0) | (df$hs_opposite1_dist_4 == 0 & df$hs == 1),1,0)
df$hs_dist5 = ifelse((df$hs_opposite1_dist_5 == 1 & df$hs == 0) | (df$hs_opposite1_dist_5 == 0 & df$hs == 1),1,0)
df$hs_dist6 = ifelse((df$hs_opposite1_dist_6 == 1 & df$hs == 0) | (df$hs_opposite1_dist_6 == 0 & df$hs == 1),1,0)
df$hs_dist7 = ifelse((df$hs_opposite1_dist_7 == 1 & df$hs == 0) | (df$hs_opposite1_dist_7 == 0 & df$hs == 1),1,0)
df$hs_dist8 = ifelse((df$hs_opposite1_dist_8 == 1 & df$hs == 0) | (df$hs_opposite1_dist_8 == 0 & df$hs == 1),1,0)
df$hs_dist9 = ifelse((df$hs_opposite1_dist_9 == 1 & df$hs == 0) | (df$hs_opposite1_dist_9 == 0 & df$hs == 1),1,0)
df$hs_dist10 = ifelse((df$hs_opposite1_dist_10 == 1 & df$hs == 0) | (df$hs_opposite1_dist_10 == 0 & df$hs == 1),1,0)


# HOMEOWNERSHIP
df$homeowner = ifelse(df$ownershp==10,1,ifelse(df$ownershp==20,0,NA))




df$employed = ifelse(df$empstat %in% 10:15,1,ifelse(df$empstat%in%20:29,0,NA))



#####################################
#### SET COVARIATES FOR FORMULAS ####
#####################################

covars = c( 
  "age", "hs","incwage", "mover","hoh_age", "homeowner",
  "famsize", 'employed',  'hrswork1', 'wkswork1',
  paste0("opposite1_dist_", 1:k), paste0("hs_dist", 1:k))



#########################
#### ESTIMATE MODELS ####
#########################


## NO FE
fml_dem = as.formula(paste(paste("democrat ~ ", paste(covars, collapse= "+")),"|0|0|county_state",sep=""))
mod_dem1 = felm(formula=fml_dem, data = df[df$white==1,] )
sum_dem1 = summary(mod_dem1, robust =T)


## STATE FE
fml_dem_st = as.formula(paste(paste("democrat ~ ", paste(covars, collapse= "+")),"|statefip|0|county_state",sep=""))
mod_dem_st = felm(formula=fml_dem_st, data = df[df$white==1,] )
sum_dem_st=summary(mod_dem_st, robust =T)


## COUNTY FE
fml_dem_c = as.formula(paste(paste("democrat ~ ", paste(covars, collapse= "+")),"|county_state|0|county_state",sep=""))
mod_dem_c = felm(formula=fml_dem_c, data = df[df$white==1,] )
sum_dem_c=summary(mod_dem_c, robust =T)


## ENUMDIST FE
fml_dem_ed = as.formula(paste(paste("democrat ~ ", paste(covars, collapse= "+")),"|cnty_enumdist|0|county_state",sep=""))
mod_dem_ed= felm(formula=fml_dem_ed, data = df[df$white==1,] )
sum_dem_ed=summary(mod_dem_ed, robust =T)



## REEL FE
fml_dem_r = as.formula(paste(paste("democrat ~ ", paste(covars, collapse= "+")),"|reel|0|county_state",sep=""))
mod_dem_r = felm(formula=fml_dem_r, data = df[df$white==1,] )
sum_dem_r=summary(mod_dem_r, robust =T)



## REEL PAGE 5 FE
fml_dem_rp5 = as.formula(paste(paste("democrat ~ ", paste(covars, collapse= "+")),"|reel_page5|0|county_state",sep=""))
mod_dem_rp5= felm(formula=fml_dem_rp5, data = df[df$white==1,] )
sum_dem_rp5=summary(mod_dem_rp5, robust =T)


## REEL PAGE 10 FE
fml_dem_rp10 = as.formula(paste(paste("democrat ~ ", paste(covars, collapse= "+")),"|reel_page10|0|county_state",sep=""))
mod_dem_rp10 = felm(formula=fml_dem_rp10, data = df[df$white==1,] )
sum_dem_rp10 = summary(mod_dem_rp10, robust =T)



################
#### FIGURE ####  
################

## MAKE PLOT DATA
plot_data = as.data.frame(matrix(c(sum_dem1$coefficients['opposite1_dist_1',c(1,2,4)], "controls",  "none"), ncol =5, nrow=1), stringsAsFactors = F)
colnames(plot_data) = c("coef", "se", "p-value", "covars", 
                        "fe")

plot_data = rbind(plot_data, c(sum_dem_st$coefficients['opposite1_dist_1', c(1,2,4)], "controls",  "state"))
plot_data = rbind(plot_data, c(sum_dem_c$coefficients['opposite1_dist_1', c(1,2,4)], "controls",  "county"))
plot_data = rbind(plot_data, c(sum_dem_r$coefficients['opposite1_dist_1', c(1,2,4)], "controls",  "reel"))

plot_data = rbind(plot_data, c(sum_dem_rp5$coefficients['opposite1_dist_1', c(1,2,4)], "controls",  "page-5"))
plot_data = rbind(plot_data, c(sum_dem_rp10$coefficients['opposite1_dist_1', c(1,2,4)], "controls",  "page-10"))
plot_data = rbind(plot_data, c(sum_dem_ed$coefficients['opposite1_dist_1', c(1,2,4)], "controls",  "enumdist"))

plot_data$coef=as.numeric(plot_data$coef)
plot_data$se=as.numeric(plot_data$se)
plot_data$`p-value`=as.numeric(plot_data$`p-value`)
plot_data$covars = factor(plot_data$covars, levels = c("no controls", "controls"))
plot_data$fe=factor(plot_data$fe, levels = c("none", "state", "county", "enumdist", "reel","page-10",
                                               "page-5"))


plot_data_0509 = plot_data


rm(list=ls()[ls()!='k' & !grepl('plot_data', ls())])
gc()


##############################
#### 2017 SAMPLE ANALYSIS ####
##############################

load("02-data/seg_analysis_2017.Rdata")
  df = as.data.table(df)
  
  
 df$white = ifelse(df$race == 100, 1, 0) # WHITE INDICATOR
  
  #### FIXED EFFECT VARIABLES
  
  # county
  df$county_state = paste0(df$statefip, "_", df$county)  
  
 
  # enumdist
  df$cnty_enumdist = paste0(df$county_state, "_", df$supdist,"_",df$enumdist)
  df$cnty_enumdist = ifelse(grepl('NA', df$cnty_enumdist), NA, df$cnty_enumdist)
  
  # 10 page
  df$page10 = round(df$pageno, digits = -1)
  df$reel_page10 = paste0(df$reel, "_", df$page10)
  
  # 5 page
  mround <- function(x,base){ 
    base*round(x/base) }
  df$page5 = mround(df$pageno, 5)
  df$reel_page5 = paste0(df$reel, "_", df$page5)
  
  
  df$hs_dist1 = ifelse((df$hs_opposite1_dist_1 == 1 & df$hs == 0) | (df$hs_opposite1_dist_1 == 0 & df$hs == 1),1,0)
  df$hs_dist2 = ifelse((df$hs_opposite1_dist_2 == 1 & df$hs == 0) | (df$hs_opposite1_dist_2 == 0 & df$hs == 1),1,0)
  df$hs_dist3 = ifelse((df$hs_opposite1_dist_3 == 1 & df$hs == 0) | (df$hs_opposite1_dist_3 == 0 & df$hs == 1),1,0)
  df$hs_dist4 = ifelse((df$hs_opposite1_dist_4 == 1 & df$hs == 0) | (df$hs_opposite1_dist_4 == 0 & df$hs == 1),1,0)
  df$hs_dist5 = ifelse((df$hs_opposite1_dist_5 == 1 & df$hs == 0) | (df$hs_opposite1_dist_5 == 0 & df$hs == 1),1,0)
  df$hs_dist6 = ifelse((df$hs_opposite1_dist_6 == 1 & df$hs == 0) | (df$hs_opposite1_dist_6 == 0 & df$hs == 1),1,0)
  df$hs_dist7 = ifelse((df$hs_opposite1_dist_7 == 1 & df$hs == 0) | (df$hs_opposite1_dist_7 == 0 & df$hs == 1),1,0)
  df$hs_dist8 = ifelse((df$hs_opposite1_dist_8 == 1 & df$hs == 0) | (df$hs_opposite1_dist_8 == 0 & df$hs == 1),1,0)
  df$hs_dist9 = ifelse((df$hs_opposite1_dist_9 == 1 & df$hs == 0) | (df$hs_opposite1_dist_9 == 0 & df$hs == 1),1,0)
  df$hs_dist10 = ifelse((df$hs_opposite1_dist_10 == 1 & df$hs == 0) | (df$hs_opposite1_dist_10 == 0 & df$hs == 1),1,0)
  
  
  
  
  
  # HOMEOWNERSHIP
  df$homeowner = ifelse(df$ownershp==10,1,ifelse(df$ownershp==20,0,NA))

  
  

  df$employed = ifelse(df$empstat %in% 10:15,1,ifelse(df$empstat%in%20:29,0,NA))
  

  
  #####################################
  #### SET COVARIATES FOR FORMULAS ####
  #####################################
  
  covars = c( 
    "age", "hs","incwage", "mover","hoh_age", "homeowner",
    "famsize", 'employed',  'hrswork1', 'wkswork1',
    paste0("opposite1_dist_", 1:k), paste0("hs_dist", 1:k))
  
  

  #########################
  #### ESTIMATE MODELS ####
  #########################
  
  
  ## NO FE
  fml_dem = as.formula(paste(paste("democrat ~ ", paste(covars, collapse= "+")),"|0|0|county_state",sep=""))
  mod_dem1 = felm(formula=fml_dem, data = df[df$white==1,] )
  sum_dem1 = summary(mod_dem1, robust =T)
  
 
  
  ## STATE FE
  fml_dem_st = as.formula(paste(paste("democrat ~ ", paste(covars, collapse= "+")),"|statefip|0|county_state",sep=""))
  mod_dem_st = felm(formula=fml_dem_st, data = df[df$white==1,] )
  sum_dem_st=summary(mod_dem_st, robust =T)
  
 
  ## COUNTY FE
  fml_dem_c = as.formula(paste(paste("democrat ~ ", paste(covars, collapse= "+")),"|county_state|0|county_state",sep=""))
  mod_dem_c = felm(formula=fml_dem_c, data = df[df$white==1,] )
  sum_dem_c=summary(mod_dem_c, robust =T)
  
 
  ## ENUMDIST FE
  fml_dem_ed = as.formula(paste(paste("democrat ~ ", paste(covars, collapse= "+")),"|cnty_enumdist|0|county_state",sep=""))
  mod_dem_ed= felm(formula=fml_dem_ed, data = df[df$white==1,] )
  sum_dem_ed=summary(mod_dem_ed, robust =T)
  
 
  
  ## REEL FE
  fml_dem_r = as.formula(paste(paste("democrat ~ ", paste(covars, collapse= "+")),"|reel|0|county_state",sep=""))
  mod_dem_r = felm(formula=fml_dem_r, data = df[df$white==1,] )
  sum_dem_r=summary(mod_dem_r, robust =T)
 
  


  
  
  ## REEL PAGE 5 FE
  fml_dem_rp5 = as.formula(paste(paste("democrat ~ ", paste(covars, collapse= "+")),"|reel_page5|0|county_state",sep=""))
  mod_dem_rp5= felm(formula=fml_dem_rp5, data = df[df$white==1,] )
  sum_dem_rp5=summary(mod_dem_rp5, robust =T)
  
 
  ## REEL PAGE 10 FE
  fml_dem_rp10 = as.formula(paste(paste("democrat ~ ", paste(covars, collapse= "+")),"|reel_page10|0|county_state",sep=""))
  mod_dem_rp10 = felm(formula=fml_dem_rp10, data = df[df$white==1,] )
  sum_dem_rp10 = summary(mod_dem_rp10, robust =T)
 
  
  
  ################
  #### FIGURE ####  
  ################
  
  ## MAKE PLOT DATA
  plot_data = as.data.frame(matrix(c(sum_dem1$coefficients['opposite1_dist_1',c(1,2,4)], "controls",  "none"), ncol =5, nrow=1), stringsAsFactors = F)
  colnames(plot_data) = c("coef", "se", "p-value", "covars", 
                          "fe")
  
  plot_data = rbind(plot_data, c(sum_dem_st$coefficients['opposite1_dist_1', c(1,2,4)], "controls",  "state"))
  plot_data = rbind(plot_data, c(sum_dem_c$coefficients['opposite1_dist_1', c(1,2,4)], "controls",  "county"))
  plot_data = rbind(plot_data, c(sum_dem_r$coefficients['opposite1_dist_1', c(1,2,4)], "controls",  "reel"))
  plot_data = rbind(plot_data, c(sum_dem_rp5$coefficients['opposite1_dist_1', c(1,2,4)], "controls",  "page-5"))
  plot_data = rbind(plot_data, c(sum_dem_rp10$coefficients['opposite1_dist_1', c(1,2,4)], "controls",  "page-10"))
  plot_data = rbind(plot_data, c(sum_dem_ed$coefficients['opposite1_dist_1', c(1,2,4)], "controls",  "enumdist"))
  
  
  plot_data$coef=as.numeric(plot_data$coef)
  plot_data$se=as.numeric(plot_data$se)
  plot_data$`p-value`=as.numeric(plot_data$`p-value`)
  plot_data$covars = factor(plot_data$covars, levels = c("no controls", "controls"))
  plot_data$fe=factor(plot_data$fe, levels = c("none", "state", "county", "enumdist", "reel","page-10",
                                               "page-5"))
  
  
  
  plot_data_17 = plot_data
  
 

rm(list=ls()[ls()!='k' & !grepl('plot_data', ls())])
gc()



############################################
#### CREATE PLOT OF BOTH SAMPLE EFFECTS ####
############################################

plot_data = plot_data_0509 %>%
  as_tibble %>%
  mutate(year = '2005/2009 Sample', Neighbor = k) %>%
  bind_rows(plot_data_17 %>%
              as_tibble %>%
              mutate(year = '2017 Sample', Neighbor = k))%>%
  mutate(fe = capitalize(as.character(fe)),
         fe = ifelse(fe=='Enumdist','District',fe),
         fe = factor(fe, levels = c('None', 'State', 'County', 'Reel', 'District', 'Page-10', 'Page-5')))



list_of_plot_data[[paste0('k',k)]] = plot_data
print(k)
}  


plot_data = rbindlist(list_of_plot_data)


  ## LOAD AESTHETIC
source("01-code/r_utils.R")
require(scales)
plot_data=plot_data%>%
  mutate(fe = capitalize(as.character(fe)),
         fe = ifelse(fe=='Enumdist','District',fe),
         fe = factor(fe, levels = c('None', 'State', 'County', 'Reel', 'District', 'Page-10', 'Page-5')))

## MAKE PLOT
p=ggplot(data = plot_data, aes(x = fe, y = coef, group=Neighbor,color = as.numeric(fe),fill = as.numeric(fe))) +
  scale_shape_manual(name = '', values = c(21, 22)) +
  scale_color_manual(name = '', values = c('black', 'grey40')) + 
  xlab('')+
  ylab('Black Neighbor Coefficient') + geom_hline(yintercept = 0, alpha = .5, linetype = "dashed")+
  geom_errorbar(aes(ymin=coef-1.96*se, ymax=coef+1.96*se), width=0, 
                position=position_dodge(.75), color = "dodgerblue", alpha=.75) + 
  geom_point(shape = 21,size = 3, position=position_dodge(.75), color = "dodgerblue", fill = "white")  +# scale_y_continuous(limits=c(-0.01,0.3),oob=squish)+
  guides(alpha=NULL, fill=NULL)+
  theme_shom()+
  theme(text = element_text(size=18), legend.position="bottom",
        legend.title = element_blank(),
        axis.text.x = element_text(angle = 45, hjust = 1))+
  #scale_y_continuous(limits=c(-0.02,0.075),oob=squish)+
  guides(color='none', fill = 'none', shape='none')+
  facet_wrap(year~.,nrow = 2, strip.position = 'top')+
  NULL
p

## SAVE PLOT
ggsave(file = "03-output/01-plots/FigS24.jpeg", plot = p, width = 9, height = 5, units = 'in', dpi=600)


